Comparative health care use patterns of people with schizophrenia near the end of life: A population-based study in Manitoba, Canada

Context: The rate of health care and palliative care utilization for patients with schizophrenia near the end-of-life is currently unknown. Objective: Compare rate of health care services, including palliative care, used in the last 6-24 months of life for patients with and without schizophrenia. Design: Using the de-identified administrative data Repository at the Manitoba Centre for Health Policy; a matched cohort study between 1995/96 and 2007/08, comparing healthcare services utilized six months to two years prior to death of all (de-identified) decedents with a diagnosis of schizophrenia > 10 years to decedents without a schizophrenia diagnosis. Setting: province of Manitoba, Canada (population 1.235 million). Participants: Schizophrenia definition:ICD-9-CM 295, or ICD-10-CA code of F20, F21, F23.2, F25 in hospital or physician files, over a 12-year period 1987-1998, in the 12 years prior to death for each individual. Decedents were matched (1:3) on age, sex, geography and date of death +/- 2 months. Main outcome measures: Health service utilization rates within six-months to two years prior to death. Results: In the last six months of life, compared to their matched cohort: decedents with schizophrenia had higher rates (52.1% vs. 24.4%, p<.00001) and number of days (89.2 vs. 40.3 days, p<.0001) residing in a nursing home; had higher ambulatory visit rates to general practitioners (6.4 vs. 5.5 visits per person, p<.0001), higher rate of visits to psychiatrists (0.53 vs. 0.07 visits per person) and lower rates of seeing other specialists. Theywere less likely to have opioid analgesia (aRR=0.7157, p-value=0.0006) or to receive palliative care (aOR=0.48, 95% CI 0.41-0.57). Conclusion: End-of-life care is lacking for patients with schizophrenia. Compared to their matched cohort, these patients were much more likely to die in nursing homes, less likely to see specialists (other than psychiatrists), less likely to be prescribed analgesics, and less likely to receive palliative care. (C) 2012 Elsevier B.V. All rights reserved.